What happened?
Applied Compute, which offers an AI cloud infrastructure that lets enterprises train and run large language models (LLMs), has raised $350 million. The round was led by Elad Gil, with participation from Kleiner Perkins, Lux Capital, Hanabi Capital, NVIDIA, Blackstone, Shopify Ventures and HubSpot Ventures.
Founded in 2025 by Yash Patil, Rhythm Garg and Linden Li, the company's valuation rose to $3.25 billion with this round. It had previously raised $80 million in October last year.
What it offers
Applied Compute's platform lets developers manage model training and inference through a single infrastructure. The company also supports customers with machine learning research.
One standout feature is its use of asynchronous reinforcement learning to speed up training: data generation and model weight updates run simultaneously. The company aims to reduce the need for teacher models in training through self-distillation.
Shift-1 and automatic scaling
The platform's Shift-1 model router directs requests from coding agents to the most cost-effective model and can distribute complex tasks across multiple models. The infrastructure also automatically scales hardware resources based on user traffic and enables regular fine-tuning so models adapt to customer needs.
Where the money goes
Applied Compute said it will use the new funding to buy more hardware to expand its market presence and strengthen its AI infrastructure.
